1. Field of the Invention
The present invention relates to a semiconductor device having a circuit comprising a thin film transistor on a substrate having an insulating surface, and a manufacturing method therefor. In particular, the present invention can ideally be used in electro-optical devices, typically liquid crystal display devices in which an active matrix circuit and a driver circuit formed on its periphery, are formed on the same substrate, and in electronic equipment loaded with an electro-optical device. Note that semiconductor device, in this specification, indicates general devices that function by using semiconductor characteristics. Note also that the above stated electro-optical devices, and electronic equipment loaded with the electro-optical device, are included in that category.
2. Description of the Related Art
The development of semiconductor devices having large surface area integrated circuits formed by thin film transistors (hereinafter referred to as TFTs) on a substrate having an insulating surface is advancing. Active matrix type liquid crystal display devices, EL display devices, and contact type image sensors are known as typical examples of such. TFTs are classified by their structure and their method of manufacture. In particular, the electric field effect mobility is high for TFTs (referred to as crystalline, TFTs) in which a semiconductor film having a crystal structure is made into an active layer, so that it is possible to form circuits with a variety of functions.
For example, a pixel section or pixel matrix circuit formed by n-channel TFTs, driver circuits such as a shift register circuit, a level shifter circuit, and a buffer circuit, based on CMOS circuits, and a sampling circuit are formed in each functional block on one substrate in an active matrix type liquid crystal display device. In addition, integrated circuits in a contact type image sensor, such as a sample hold circuit, a shift register circuit, and a multiplexer circuit, are formed using TFTs.
The characteristics of an electric field effect transistor such as a TFT can be considered to be divided into a linear region in which the drain current and the drain voltage increase proportionally, a saturation region in which the drain current is saturated even if the drain voltage increases, and a cut-off region in which ideally current does not flow even if there is an applied drain voltage. The linear region and the saturation region are called the ON region of a TFT, while the cut-off region is called the OFF region in this specification. In addition, for convenience the drain current in the ON region is called the ON current, and the current in the OFF region is called the OFF current.
The operating conditions of the respective circuits are not necessarily identical, so that naturally the characteristics required in the TFT also differ a great deal. In the pixel section, there is a structure formed by an n-channel TFT switching element and an auxiliary storage capacitor, and this is driven by applying a voltage to the liquid crystal. It is necessary to drive the liquid crystal by an alternating current here, and a system called frame inversion driving is employed. Therefore, a required TFT characteristic is the necessity to sufficiently reduce the leakage current. In addition, a high drive voltage is applied to the buffer circuit, so that it is necessary to increase the voltage resistance. Furthermore, it is necessary to sufficiently maintain the ON current in order to increase the current driver performance.
However, there is a problem in that the off current of the crystalline TFT is liable to become large. From the point of reliability, it is still believed that the crystalline TFT fall short of a MOS transistor (a transistor manufactured on a single crystal semiconductor substrate) used in LSIs, etc. For example, a deterioration phenomenon of a drop in the ON current in the crystalline TFT has been observed. The cause of this is the hot carrier effect, and it is thought that the hot carrier generated by the high electric field in the vicinity of the drain causes the degradation phenomenon.
A lightly doped drain (LDD) structure is known in a TFT structure. This structure is formed by a low concentration impurity region between a channel region, and a source region or drain region in which a high concentration of impurities is doped. This low concentration impurity region is called an LDD region. In addition, for the LDD structure, depending upon the positional relationship with the gate electrode, there is an LDD structure that overlaps the gate electrode (hereinafter, this LDD structure is referred to as GOLD (gate-drain overlapped LDD)), and an LDD structure that does not overlap the gate electrode. The high electric field is eased, the hot carrier effect is prevented, and the reliability can be increased with a GOLD structure. For example, there is a GOLD structure in which sidewalls are formed by silicon in Mutsuko Hatano, Hajime Akimoto and Takeshi Sakai, IEDM97 Technical Digest, pp. 523-6, 1997”, and compared to TFTs with other structures, it has been confirmed that a very superior reliability can be obtained.
In addition, there is a TFT placed in each of from several tens to several millions of pixels in the pixel section of the active matrix type liquid crystal display device, and a pixel electrode is formed in each of the TFTs. Opposing electrodes are formed on the side of the opposing substrate sandwiching the liquid crystal, forming a kind of capacitor with the liquid crystal as a dielectric. The electric potential applied to each pixel is then controlled by the TFT switching function, and this becomes a structure in which the liquid crystals are driven by controlling the electric charge to the capacitors, controlling the amount of light transmitted and displaying an image.
The capacity of this capacitor gradually decreases due to the leak current, so that this causes the amount of transmitted light to change and the contrast of the image display to be reduced. Capacitor lines are formed conventionally, and a separate capacitor (a storage capacitor) is formed in parallel to the capacitor with the liquid crystal as its dielectric. The storage capacitor works to supplement the capacity lost by the capacitor with the liquid crystal as its dielectric.
However, the required characteristics are not necessarily the same for a TFT as a pixel section switching element and a driver circuit TFT such as a shift register circuit or a buffer circuit. For example, a large inverse bias voltage (negative for an n-channel TFT) is applied to the gate electrode in the pixel section TFT, but there is basically no operation in which an inverse bias voltage is applied to the driver circuit TFT. In addition, the operation speed of the former may be less than 1/100 that of the latter. Thus it is not preferable to use a similar structure for TFT in which the operating condition and required characteristics differ largely.
Furthermore, compared with an ordinary LDD structure, there is a problem with the GOLD structure in that the OFF current becomes large. In order to prevent an increase in the OFF current, it is possible to make a multi-gate structure in which a plural number of gates are formed between one source and drain pair, but that is insufficient for the GOLD structure TFT. Therefore, it is not necessarily preferable to form all of the TFTs of a large surface area integrated circuit with the same structure. For example, with the n-channel TFT constituting the pixel section, if the OFF current increases, then the power consumption increases and abnormalities in the image display appear, so that it is not desirable to apply the GOLD structure crystalline TFT as is. In addition, there is a problem with the LDD structure that has no overlap with the gate electrode in that the ON current decreases due to an increase in the series resistance. The ON current can be freely designed by the channel width, and for example, it is not always necessary to form the LDD structure that does not overlap the gate electrode, in a TFT constituting a buffer circuit.
In addition, if a storage capacitor using capacitor wirings in the pixel section is formed to maintain a sufficient capacity, then the aperture ratio must be sacrificed. In particular, for a small size high definition panel used in a projector type display device, the pixel area for each pixel is also small, so that the reduction in the aperture ratio due to the capacitor wiring becomes a problem.